US 8027671 B2
Among other things, in controlling a download of one or more files from a server to a mobile device, account is taken of at least two of: an urgency of the file, the existence of a user-indicated preference about the download, a power status of the mobile device, and a network connectivity status of the mobile device.
1. A method comprising:
downloading a file over a communication channel from a server to a mobile device,
monitoring, in the course of the download, a first condition comprising a capacity of the communication channel,
monitoring, in the course of the download, a second condition comprising a state of resources at the mobile device that will be needed for the download,
monitoring, in the course of the download, a third condition comprising an urgency of the file,
continuing a partial download of the file, even when one or more of the conditions has deteriorated, if predetermined circumstances exist,
suspending a download of the file that is occurring when a power level or network connectivity degrades during the download, and
resuming the suspended download when the power level or network connectivity improves during the suspension.
2. The method of
3. The method of
4. The method of
5. The method of
6. The method of
enabling a user of the mobile device, to which large files are normally downloaded automatically from a server based on the user's preferences, to initiate a download by invoking a ‘sync now’ option on a user interface of the mobile device.
7. The method of
at the mobile device, receiving a sync signal from the server with respect to one or more large files to be downloaded, and
responding to the sync signal by requesting and accepting a download of the one or more large files.
8. The method of
at the mobile device, displaying to a user an indication of the time when a download of one or more large files to the mobile device is expected to occur.
9. The method of
adjusting a level of usage of the communication channel to download one or more large files to the mobile device based on a time of day when the download is to occur.
10. The method of
11. The method of
sending from the mobile device to the server, a log of download activity with respect to one or more large files that are downloaded to the mobile device.
12. The method of
enabling a user of the mobile device to specify one or more types of files that are to be downloaded to the mobile device, and storing the specification of file types in a server in association with an identification of the mobile device or the user.
13. The method of
14. The method of
displaying to a user of the mobile device a list of files that are in process of being downloaded to the device from the server, but which have not yet been completely downloaded to the device.
15. The method of 14 in which the displaying includes showing a percentage of at least one of the listed files which has been downloaded.
16. The method of 14 in which the displaying occurs at the mobile device.
17. The method of 14 in which the displaying occurs on a web page.
This description relates to delivering files to a mobile device.
As shown in
In the email service provided on a RIM Blackberry device, for example, once the software on the mobile device is set up, email messages automatically arrive when they are available, with no user action required.
In general, in an aspect, in controlling a download of one or more files from a server to a mobile device, account is taken of at least two of: a power status of the mobile device, a network connectivity status of the mobile device, an urgency of the file, and the existence of a user-indicated preference about the download.
Some implementations include one or more of the following features. The controlling includes suspending a download of a file that is occurring when a power level or network connectivity degrades during the download. The suspending of the download is based on a set of rules or heuristics. The suspended download is resumed when the power level or network connectivity improves during the suspension. The resumed download does not include at least some of the file or files that were downloaded before the suspension. The resumed download does not include one or more pieces of a file that were downloaded before the suspension.
In general, in an aspect, download of a file from a server to a mobile device through a network is controlled by breaking the file into pieces to be sent separately, monitoring conditions of the mobile device and the network with respect to each piece, considering suspending download of a subsequent piece if the conditions deteriorate with respect to a piece.
Some implementations include one or more of the following features. The conditions are monitored after each piece is sent and before download of the subsequent piece has begun. A partial download of a file is allowed to continue under deteriorated conditions, if predetermined circumstances exist.
In general, in an aspect, a user of a mobile device, to which large files are normally downloaded automatically from a server based on the user's preferences, can initiate a download by invoking a ‘sync now’ option on a user interface of the mobile device.
In general, in an aspect, at a mobile device, a sync signal is received from a server with respect to one or more large files to be downloaded, and to the sync signal is responded to by requesting and accepting a download of the one or more large files.
In general, in an aspect, at a mobile device, an indication is displayed to a user of the time when a download of one or more large files to the mobile device is expected to occur.
In general, in an aspect, an amount of bandwidth used to download one or more large files to a mobile device is adjusted based on a time of day when the download is to occur.
Some implementations include one or more of the following features. The amount of bandwidth used at a given time of the day is governed by an amount paid by a user of the mobile device.
In general, in an aspect, a log of download activity, with respect to one or more large files that are downloaded to the mobile device, is sent from a mobile device to a server.
In general, in an aspect, a user of a mobile device can specify one or more types of files that are to be downloaded to the mobile device, and the specification of file types is stored in a server in association with an identification of the mobile device or the user.
In general, in an aspect, at a server, one or more files are selected to be downloaded to a mobile device based on stored preferences of a user of the mobile device and stored information about files that were previously downloaded to the user's device.
In general, in an aspect, a list of files that are in process of being transmitted to a client mobile device, but which have not yet been completely transmitted to the device, is displayed to a user of the device
Implementations may include one or more of the following features. The displaying includes showing a percentage of at least one of the listed the files which has been transmitted. The displaying can occur at the mobile device. The displaying can occur on a web page.
These and other features and aspects, and combinations may also be expressed as methods, apparatus, systems, program products, databases, means for performing functions, and in other ways.
Other advantages and features will become apparent from the following description and from the claims.
Some of the files 10, 20 in
In the technique described here, the user of the mobile device never needs to take any explicit action to query for new files on the server, nor to fetch the files from the server. Delivery 40 of the new files (for example, over an IP network) can occur automatically and without explicit user action or knowledge. There is no need, for example, to press any key on the mobile device or to attach a cable to the device to cause the file delivery to occur.
In contrast, typical files delivered to mobile devices today, such as email messages, audio files, and individual web pages, are typically in the 10 KB to 200 KB range each.
Delivering a file that is tens or hundreds of megabytes to a mobile device presents challenges that are absent in delivering smaller files. With large files, a complete download may take tens of minutes or even hours, depending on the speed of the connection. During this time, conditions may change. For example, at the start of the download, the mobile device may be within an area with excellent wireless network coverage; but as the download progresses, the network conditions may degrade.
Also, as the download progresses, the device's battery may lose its charge to a level at which continuing the file downloading risks depleting the battery. Generally speaking, small-file applications (music downloads, email, web browser) are not and need not be sensitive to these issues. In other words, downloading a file that is large may entail a time delay that implicates one or more other factors in the strategy used to determine the manner and timing of the download.
In the technique described here, an application 60 manages the delivery of these large files from the remote server to the mobile device in a way that is sensitive to the following considerations:
Network. The mobile device has one or more network interfaces 80, e.g., GPRS/EDGE, CDMA/EVDO, WiFi, Bluetooth, or USB cable tethered to a host computer, each of which can take advantage of a network resource for connecting to the Internet 40 (or other communication network). This connection is the “last hop” for data traveling over the IP network from the server to the device. The availability of one or more of these network resources will vary from time to time as, for example, the mobile device moves in and out of network “hotspots.” For example, at one moment, WiFi and EDGE may be available; a few minutes later, only EDGE may be available.
Power. The mobile device's power status 70 varies from time to time, depending, for example, on whether the device is charging or how long it has been on battery.
Timeliness. Certain files 10 are time sensitive and may need to be delivered right away (e.g., a stock ticker update). Other files 20 are relatively time insensitive, and a delay of hours in delivery may be acceptable (e.g., a movie trailer). Each file stored on the remote server may be annotated with an “urgency” parameter 30, which may be assigned by a creator of the file (content) or determined based on some other policy.
These considerations are sometimes interrelated. For example, a mobile device attempting to receive data over a poor-quality network connection can cause a significant drain on its battery.
We shall denote by “sync'ing” the process of copying a file from the remote server to the mobile device.
We describe a method for sync'ing that is sensitive to the above considerations. The method takes into account a set of factors (two or more) in determining a sync'ing policy, by which we mean whether and how often to sync. In some implementations, there is a particular algorithm used to take into account these factors to implement a sync'ing policy.
As shown in
The client may be installed by the device manufacturer (OEM) before delivery to the retail outlet which sells or delivers the device (we sometimes refer to the mobile device simply as a device) to the end-user. Alternatively, the client may be installed by the end-user after he has acquired the device, by downloading the software application from the Internet.
In some implementations, the client behaves as follows:
The application performs these steps transparently to the user. That is, the user need not take any action to cause the steps to be performed and the user need not be made aware of any of the steps occurring. New files are delivered—hourly, daily, weekly—directly to the mobile device. No user action is required. This is similar to the behavior of a home Digital Video Recorder (DVR).
Once the files have been downloaded, they can be stored on the device's file system. For example, on a Windows Mobile device, the files could be stored in the “My Documents” folder. The user can access them manually, using existing tools on the mobile device (e.g., a “File Manager” application on a Windows Mobile device).
However, because user-access to a device file system is not always available (e.g., on many Java-based phones) and because even when such access is available it is not always intuitive, the client may include an application (
The delivered files may be of various types. For example, web pages (e.g., HTML), audio files (e.g., MP3), video files (e.g., h.264 or WMv9 or Flash), and so on. The application need not itself provide rendering/playout/presentation for any or all of these file types. Instead, the application may interface with existing applications (web browser, media player, etc.) on the device through standard application programming interface (API) calls. Once the user selects a file from the list (or from the native file browser), the appropriate external application is launched to render the file (
The server's behavior need not be a synchronous step-by-step process, but rather can be a set of tasks that occur asynchronously, as the server is contacted by various mobile devices 275 and web-based terminals 285. The server's behavior in some implementations includes the following (with reference again to
When a mobile device 275 queries the sync module 220 about the presence of new files, the sync module performs a multi-step process to determine if any such files are present. In some implementations, that process includes the following:
In some implementations, the sync'ing protocol includes the following features.
Account for Current Conditions
Instead of performing only a one-time check (to decide whether to sync or not) prior to a download and then downloading one or more files, the protocol also checks the device's conditions (power, network) during the download. To accomplish this, in some examples, the file 900 (in
The protocol also allows for the client to continue downloading for some limited period of time even if conditions have deteriorated.
The protocol allows for a sync (we sometimes refer to a sync'ing procedure simply as a sync) to be initiated in several ways.
One way is to use a timer on the client; when the timer expires, the client queries the server for the presence of new files. In the sample algorithm set forth below, the proposed counter starts at 300 seconds. To perform the count down, the client application must be always running on the device. This is much like a standard mobile application (e.g., a game), which is either active or closed, and when closed, does not consume any random-access memory on the device. In these implementation, the client application must be auto-started when the device is booted, and it must persist while the device is on. This is analogous to a home DVR, which is always on.
A second way to initiate a sync is in response to a user request. The client application may provide a “Sync Now” button or link. (
(1) Attempt to perform an immediate sync, regardless of the conditions.
(2) Evaluate the ‘downloadAllFiles’ variable in the below algorithm, based on current conditions. If the variable evaluates to ‘true’, then the sync is performed. If not, the cause for failure may be presented to the user with a prompt, e.g., “The battery is only ⅓ full. Sync'ing may drain the battery further and risk draining it altogether. Do you wish to sync anyway?”
A third way to initiate a sync is from the server. In some settings, the client may be capable of receiving asynchronous messages directly from the server. For example, a RIM Blackberry device provisioned to use a Blackberry Enterprise Server (BES) can receive signals directly from the server through an MDS component. Therefore, the client application need not continuously poll the server for new files, nor remain active. The application can, instead, be inactive until the server signal arrives and causes the device to launch the application.
The features described above (and other features in various combinations) can be implemented in a wide variety of algorithms. Among other things, different algorithms may specify different techniques for suspending (and subsequently resuming) file downloads as conditions change. We provide one example of an algorithm below. We first introduce several terms and parameters that are relevant to the example.
The example algorithm follows:
Voice and SMS traffic generate more revenue per bit to the owner of a mobile network than generic data traffic. Therefore, during daytime hours (e.g., 9 AM-7 PM), when voice and SMS traffic is high and spare capacity limited, network owners are likely to discourage network-intensive applications. On the other hand, most cellular networks are ‘latent’ during off-peak hours such as 9 PM-6 AM. During this time window, mobile network owners are likely to be more receptive to network-intensive applications.
In short, bandwidth on a wireless radio-access network (e.g., EDGE or EVDO) is a commodity whose value to the network owner varies over time.
The application can accommodate this constraint by disabling sync activity altogether during daytime hours when sync'ing requires mobile network access. The result is that users may only receive updates during the evening hours.
Alternatively, the application can throttle its mobile network bandwidth usage to an acceptable level during the daytime hours by waiting a specified period of time between downloading chunks of data. For example, the application can cut by 50% its bandwidth usage by pausing for ‘timeForThisChunk’ seconds after the chunk (e.g., a piece of a file) has been downloaded. Similarly, the application can cut by 67% its bandwidth usage by pausing for twice this amount of time.
A network operator can offer a tiered level of service to its customers, in which those customers who pay for a ‘premium service’ are permitted higher throughput for during-the-day downloads over the mobile network. The application used by premium customers would not be forced to pause (as long, or at all) between chunk downloads.
Other implementations are also within the scope of the following claims.